As the Innocence Project reports, Texas executed him in Dallas in the 1950s. He was a young Black man. Decades later, a court reviewed his case and declared him innocent, almost 70 years after the state killed him. His name is now cleared on paper. His life is still gone.

For me, that story lands like a system failure report. Not a glitch. A fatal error.

When a system harms one person in this way, the next question is not “Was this rare.” The next question is “How often does this happen, and can the system’s error rate ever be low enough for an irreversible punishment.” Once you dig into the numbers on the death penalty, the answer is clear. No.

These are error rates that NO engineer would accept.

Researchers who study wrongful convictions have tried to estimate how many people on death row were innocent. A well-known 2014 study in the journal PNAS estimated that at least about 4 percent of people sentenced to death in the modern era were innocent. That estimate came from survival analysis methods that treated exonerations as “events” and controlled for people leaving death row for other reasons.

Four percent sounds small at first glance. It is not small.

Imagine a medical device with a four percent chance of killing the wrong patient. No hospital would approve it. Imagine a plane crash rate of four per hundred flights. No one would board.

Now turn back to the death penalty. Since the 1970s, more than 190 people sentenced to death in the United States have been exonerated and released from death row. That is the group we know about. Each case is a Tommy Lee Walker story with a slightly different outline. New evidence. DNA. Witness recantations. Misconduct revealed. Those exonerations are not the whole set of errors. They are the visible tip.

From a data view, the pattern looks like this:

• Non-trivial error rate.
• Irreversible outcome.
• Strong evidence that some errors are never caught.

For any other system, that combination would lead to one conclusion. Shut this down.

So who pays for this?

Carceral capitalism is the idea that punishment has become tied to profit and political gain. The death penalty sits inside that system. It does not stand apart as a pure moral statement. It has budgets, vendors, risks, and incentives like any large program. Capital cases cost more than non-capital ones at every stage. Studies from several states show that prosecutions that seek death run hundreds of thousands, sometimes millions, of dollars higher than cases that seek life without parole. Extra cost comes from longer trials, more lawyers, more experts, and years of appeals.

The public does not see a return on that investment in safety. States that use the death penalty do not show lower murder rates than states that have abandoned it. In some years, states without the death penalty have had lower homicide rates. That pattern repeats often enough that you cannot claim a safety gain.

So what are we buying.

We are buying an expensive symbol. District attorneys win elections on “tough on crime” records. Some judges build careers on their trial results. Vendors get contracts for expert testimony, specialized defense, and death row facilities. Taxpayers cover the bills. Families of victims carry the grief either way. From a carceral capitalism lens, the death penalty looks like a high-cost, low-value product. It absorbs money that could fund things that actually cut violence, like housing, mental health care, or trauma-informed programs. It does not deliver measurable public safety benefits.

Supporters of the death penalty often say it prevents murder. The logic sounds clean: if people fear execution, they will avoid killing. But you have to test claims like that with data, not with intuition. The National Research Council examined decades of deterrence studies on the death penalty. The panel found that those studies did not give reliable evidence that the death penalty reduces homicides compared with long prison sentences. Problems included weak models, missing factors, and unstable results across methods. When you adjust for those issues, the claimed “deterrent effect” falls apart.

So the best summary from that review is simple. We do not have solid evidence that the death penalty saves lives. We do have solid evidence that it ends lives for sure. Some of those lives are innocent.

And what is behind all of this?

If you map death sentences and executions, race shows up over and over again.

Black people make up around a tenth of the U.S. population but a much larger share of people on death row. In many states, cases with white victims are far more likely to result in death sentences than cases with Black victims. Studies that control for other factors still find higher odds of a death sentence when the victim is white, especially when the defendant is Black.

Those are not random quirks. They are the traces of whose lives power has valued more over time.

The Tommy Lee Walker case fits that pattern. A young Black man in the Jim Crow South. Limited legal defense. A system ready to accept thin evidence and move fast toward execution. His posthumous exoneration shows that race and power can shape not just who gets arrested, but who gets killed by the state. From a data standpoint, this is not a fair test of guilt. It is a biased sampling process. When you combine state power, racial bias, and an irreversible punishment, you do not get justice. You get structured harm.

We are using the wrong risk model.

Think about the death penalty as a kind of risk model. The state claims it can sort people into two groups. One group is so dangerous and so guilty that it deserves death. The other group does not.

For that model to be defensible, two things must be true.

1. The error rate must be close to zero.
2. The errors must be fixable.

Neither condition holds. The error rate is not near zero. We have proof of that in exoneration data and in cases like Tommy Lee Walker. The errors are not fixable once the state carries out the sentence. A later court ruling can change the paperwork. It cannot restart a heart. In most fields, we choose penalties that match the risk of error. If a model decides who receives a coupon, we accept more noise. If a model decides who gets a harsh penalty, we demand better proof and we leave room to repair mistakes. The death penalty breaks that rule. It sets the harshest penalty in the system on top of a process that still misfires.

So who is carrying the burden of proof?

The usual way the death penalty is defended goes like this. Supporters say the system is not perfect but tries hard. They point to layers of appeal and sometimes to reforms like better defense funding. They frame wrongful executions as rare exceptions. That frame puts the burden on people who oppose the death penalty. It asks them to prove that the system fails often enough to justify abolition.

A data lens flips that burden. If a state wants the power to end a life, the state should prove that its process is accurate and unbiased. It should show that no innocent person will die. It should show that the policy saves lives or prevents harm in ways no other punishment can. It cannot.

We have:

• Clear evidence of wrongful convictions in capital cases.
• Persistent racial and geographic disparities.
• No strong proof of a unique public safety benefit.
• High financial cost that pulls money away from services that prevent violence.

That package would not pass review in any other high-stakes system.

So why should EVERYONE be against it?

Some people oppose the death penalty on moral or spiritual grounds. Others see it through lived experience, after watching a loved one face a death sentence or a lifetime of appeals. Those reasons matter. A data-driven view gives another layer. You do not need shared faith or shared politics to see the pattern. The numbers show a punishment that is arbitrary, expensive, racially skewed, and prone to unrecoverable error.

Tommy Lee Walker’s posthumous exoneration is not just a tragic story from another era. It is a warning flag on the present. It tells us that once a system has proven it can take an innocent life, and once the data show that error and bias remain, the only safe rate for an irreversible punishment is zero.

The death penalty cannot reach that rate.

So if we care about fairness, if we care about public safety, if we care about not repeating the same state-made tragedies under a new date, we should take the clear step that the data support.

End the death penalty.

Sources:

In a Historic Resolution, Tommy Lee Walker Is Declared Innocent 70 Years After His Execution in Dallas

Death Penalty Information Center. Innocence Database

Gross, Samuel R., et al. “Rate of false conviction of criminal defendants who are sentenced to death.” Proceedings of the National Academy of Sciences, 2014.

National Research Council. Deterrence and the Death Penalty. National Academies Press, 2012